Update app.py

app.py

@@ -90,7 +90,7 @@ def calculate(image_in, audio_in):
     torchaudio.save("/content/audio.wav", waveform, sample_rate, encoding="PCM_S", bits_per_sample=16)
     image = Image.open(image_in)
     image = pad_image(image)
-    os.system(f"rm -rf /content/image.png")
+    # os.system(f"rm -rf /content/image.png")
     image.save("image.png")
 
     pocketsphinx_run = subprocess.run(['pocketsphinx', '-phone_align', 'yes', 'single', '/content/audio.wav'], check=True, capture_output=True)
@@ -141,8 +141,9 @@ def one_shot_talking(image_in,audio_in):
   #Improve quality of input image
   os.system(f"python /content/GFPGAN/inference_gfpgan.py --upscale 2 -i /content/image_pre.png -o /content/results  --bg_upsampler realesrgan")
   # time.sleep(60)
-  image_in_one_shot='/content/results/restored_imgs/image_pre.png'
   os.system(f"rm -rf /content/results/restored_imgs/image_pre.png")
+  image_in_one_shot='/content/results/restored_imgs/image_pre.png'
+  
   
   #One Shot Talking Face algorithm
   calculate(image_in_one_shot,audio_in)